Re: [core] WG Last Call on draft-ietf-core-new-block

["Christian M. Amsüss" <[e-mail]>]

Picking a quick one because I don't know when I'll get to the rest:

> > * 4 the new 4.08 format: Since this is a CBOR sequence, the
> >   implementation note on telling the array length in advance does not
> >   apply. As for the CDDL, I don't know whether the array format is OK to
> >   use for the sequence, or whether the CBOR
> 
> [Jon] Not sure what you are trying to ask here.  I know we removed the
> Request-Tag from the CDDL which may have changed things.

If the array were sent in CBOR, indeed the sequence 1, 10, 20 would look
like this (courtesy of cbor.me):

83    # array(3) <--- here is where you need to know the length in advance
   01 # unsigned(1)
   0A # unsigned(10)
   14 # unsigned(20)

But as the content format is a cbor-sequence one, what you actually send
is this:

   01 # unsigned(1)
   0A # unsigned(10)
   14 # unsigned(20)

(No array header, just a sequence of self-delimited items).

So now the implementer's worst worry is that if they start writing an
item in the last byte and the item needs a u16 they gotta roll back, but
that is simple. Picking the array length in advance would be hard (and
the implementation note's suggestion of not exceeding 23 items would
apply), but by using a CBOR sequence you're already free of that.

BR
c

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom